Conference PaperPDF Available

Automatic Counting and Classification of Microplastic Particles

  • January 2018
DOI:10.5220/0006725006460652
  • Conference: 7th International Conference on Pattern Recognition Applications and Methods (ICPRAM 2018)
Authors:
Javier Lorenzo-Navarro at Universidad de Las Palmas de Gran Canaria
Javier Lorenzo-Navarro
Modesto Castrillón Santana at Universidad de Las Palmas de Gran Canaria
Modesto Castrillón Santana
May Gómez at Universidad de Las Palmas de Gran Canaria
May Gómez
Alicia Herrera at Universidad de Las Palmas de Gran Canaria
Alicia Herrera
Show all 5 authorsHide
Download full-text PDFDownload full-text PDF
Read full-text
Download citation

Abstract and Figures

Microplastic particles have become an important ecological problem due to the huge amount of plastics debris that ends up in the sea. An additional impact is the ingestion of microplastics by marine species, and thus microplastics enter into the food chain with unpredictable effects on humans. In addition to the exploration of their presence in fishes, researchers are studying the presence of microplastics in coastal areas. The workload is therefore time consuming, due to the need to carry out regular campaigns to quantify their presence in the samples. So, in this work a method for automatic counting and classifying microplastic particles is presented. To the best of our knowledge, this is the first proposal to address this challenging problem. The method makes use of Computer Vision techniques for analyzing the acquired images of the samples; and Machine Learning techniques to develop accurate classifiers of the different types of microplastic particles that are considered. The obtained results show that making use of color based and shape based features along with a Random Forest classifier, an accuracy of 96.6% is achieved recognizing four types of particles: pellets, fragments, tar and line.
Figures - uploaded by Alicia Herrera
Author content
All figure content in this area was uploaded by Alicia Herrera
Content may be subject to copyright.
C105_ICPRAM
18.pdf
Content uploaded by Modesto Castrillón Santana
Author content
All content in this area was uploaded by Modesto Castrillón Santana on Mar 20, 2018
Content may be subject to copyright.
ICPRAM_20
18.pdf
Content uploaded by Alicia Herrera
Author content
All content in this area was uploaded by Alicia Herrera on Jan 28, 2018
Content may be subject to copyright.
... In microplastics research, only a few studies have tried to develop an automatic image analysis-based identification method to count and classify microplastic particles [9,[11][12][13]. Some of them [14,15] used a ZooScan based system [16], in which digital images were post-processed with the Zooprocess and Plankton Identifier software, based on the ImageJ macro language [17,18], with the ability to attribute some morphological parameters to each object counted by the software including microplastics (e.g., the maximal distance between any two points along the boundary of the object, surface area). ...
... In recent years, the use of Deep Learning approaches for object classification [20][21][22][23] exhibited performance in complex tasks like never before [11]. ...
... Before processing the image, we identified the threshold value to obtain the best possible results by contrasting the particles with the background as much as possible [11,12,32]. We considered, for the counting, all particles with a value higher than the threshold identified. ...
A Handy Open-Source Application Based on Computer Vision and Machine Learning Algorithms to Count and Classify Microplastics
Article
Full-text available
  • Jul 2021
Microplastics have recently been discovered as remarkable contaminants of all environmental matrices. Their quantification and characterisation require lengthy and laborious analytical procedures that make this aspect of microplastics research a critical issue. In light of this, in this work, we developed a Computer Vision and Machine-Learning-based system able to count and classify microplastics quickly and automatically in four morphology and size categories, avoiding manual steps. Firstly, an early machine learning algorithm was created to count and classify microplastics. Secondly, a supervised (k-nearest neighbours) and an unsupervised classification were developed to determine microplastic quantities and properties and discover hidden information. The machine learning algorithm showed promising results regarding the counting process and classification in sizes; it needs further improvements in visual class classification. Similarly, the supervised classification demonstrated satisfactory results with accuracy always greater than 0.9. On the other hand, the unsupervised classification discovered the probable underestimation of some microplastic shape categories due to the sampling methodology used, resulting in a useful tool for bringing out non-detectable information by traditional research approaches adopted in microplastic studies. In conclusion, the proposed application offers a reliable automated approach for microplastic quantification based on counts of particles captured in a picture, size distribution, and morphology, with considerable prospects in method standardisation.
View
... For detailed information, there are the most cited authors based on some important indicators, containing Total Publications, Total Cited, Links, and Total Link Strength by setting the threshold of 2 in Figure 15. Based on the results of Figure 15, in these involved authors Ferraro, Pietro, Bianco, and Vittorio are the most active through 6 publications and 41 citations from the same institute called Consiglio Nazionale delle Ricerche (CNR).Ferraro, Pietro and Bianco, Vittorio focused on researching the microplastics including microplastics identification, classifying and automatic detection via holographic imaging and machine Learning[45][46][47]. ...
A Critical Review on Artificial Intelligence—Based Microplastics Imaging Technology: Recent Advances, Hot-Spots and Challenges
Article
Full-text available
Due to the rapid artificial intelligence technology progress and innovation in various fields, this research aims to use science mapping tools to comprehensively and objectively analyze recent advances, hot-spots, and challenges in artificial intelligence-based microplastic-imaging field from the Web of Science (2019–2022). By text mining and visualization in the scientific literature we emphasized some opportunities to bring forward further explication and analysis by (i) exploring efficient and low-cost automatic quantification methods in the appearance properties of microplastics, such as shape, size, volume, and topology, (ii) investigating microplastics water-soluble synthetic polymers and interaction with other soil and water ecology environments via artificial intelligence technologies, (iii) advancing efficient artificial intelligence algorithms and models, even including intelligent robot technology, (iv) seeking to create and share robust data sets, such as spectral libraries and toxicity database and co-operation mechanism, (v) optimizing the existing deep learning models based on the readily available data set to balance the related algorithm performance and interpretability, (vi) facilitating Unmanned Aerial Vehicle technology coupled with artificial intelligence technologies and data sets in the mass quantities of microplastics. Our major findings were that the research of artificial intelligence methods to revolutionize environmental science was progressing toward multiple cross-cutting areas, dramatically increasing aspects of the ecology of plastisphere, microplastics toxicity, rapid identification, and volume assessment of microplastics. The above findings can not only determine the characteristics and track of scientific development, but also help to find suitable research opportunities to carry out more in-depth research with many problems remaining.
View
... Nowadays, over 12.7 million tons of plastic per year enter the marine environment (Haward, 2018), a figure that increased dramatically from the 2015 estimates (Jambeck et al., 2015;Lorenzo-Navarro et al., 2018). Some products of manufactured raw plastic materials like microbeads, scrubber pellets as well as materials obtained from the mechanical, photo (oxidative) and/or biological degradation of plastic form a significant number of debris and fragments of plastics, called microplastics (MPs). ...
Intelligent polarization-sensitive holographic flow-cytometer: Towards specificity in classifying natural and microplastic fibers
Article
Micron size fiber fragments (MFFs), both natural and synthetic, are ubiquitous in our life, especially in textile clothes, being necessary in modern society. In the Earth's aquatic ecosystem, microplastic fibers account for ~91% of microplastic pollution, thus deserving notable attention as one of the most alarming ecological problems. Accurate automatic identification of MFFs discharges in specific upstream locations is highly demanded. Computational microscopy based on Digital Holography (DH) and machine learning has been demonstrated to identify microplastics in respect to microalgae. However, DH is a non-specific optical tool, meaning it cannot distinguish different types of plastic materials. On the other hand, materials-specific assessments are pivotal to establish the environmental impact of different textile products and production processes. Spectroscopic assays can be employed to identify microplastics for their intrinsic specificity, although they are generally low-throughput and require large concentrations to enable effective measurements. Conversely, MFFs are usually finely dispersed within a water sample. Here we rely on a polarization-resolved holographic flow cytometer in a Lab-on-Chip (LoC) platform for analysing MFFs. We demonstrate that two important objectives can be achieved, i.e. adding material specificity through polarization analysis while operating in a microfluidic stream modality. Through a machine learning numerical pipeline, natural fibers (i.e. cotton and wool) can be clearly separated from synthetic microfilaments, namely PA6, PA6.6, PET, PP. Moreover, the proposed system can accurately distinguish between different polymers under investigation, thus fulfilling the specificity goal. We extract and select different features from amplitude, phase and birefringence maps retrieved from the digital holograms. These are shown to typify MFFs without the need for sample pre-treatment or large concentrations. The simplicity of the DH method for identifying MFFs in LoC-based flow cytometers could promote the use of polarization resolved field-portable analysis systems suitable for studying pollution caused by washing processes of synthetic textiles.
View
... In the case of particles with an identification hit quality between 700 and 300, the absorption bands were manually checked for polymer identification according to the criteria of Jung et al. (2018). Spectra of particles with a hit quality less than 300, as a limit for satisfactory identification (Primpke et al., 2017;Lorenzo-Navarro et al., 2018;Primpke et al., 2018), were compared with spectra databases for natural materials provided by Spectragryph (Kimmel_Center: Collection of 363 FTIR absorbance of natural and biogenic material of archeological interest; provided by S. Weiner from Kimmel Center for Archaeological Science, Weizmann Institute of Science, Israel). In the case of no match, the spectra were finally matched with the database of "Open Specy" (Cowger et al., 2020). ...
Deposition and in-situ translocation of microplastics in floodplain soils
Article
The microplastic (MP) contamination of oceans, freshwaters, and soils has become one of the major challenges within the Anthropocene. MP is transported in large quantities through river systems from land to sea and is deposited in river sediments and floodplains. As part of the river system, floodplains and their soils are known for their sink function with respect to sediments, nutrients, and pollutants. However, the questions remain: To what extend does this deposition occur in floodplain soils? Which spatial distribution of MP accumulations, resulting from possible environmental drivers, can be found? The present study analyzes the spatial distribution of large (L-MP, 2000–1000 μm) and medium (M-MP, 1000–500 μm) MP particles in floodplain soils of the Lahn River (Germany). Based on a geospatial sampling concept, the MP contents in floodplain soils are investigated down to a depth of 2 m through a combined method approach, including MP analyses, soil surveys, properties, and sediment dating. The analysis of the plastic particles was carried out by density separation, visual fluorescence identification, and ATR-FTIR analysis. In addition, grain-size analyses and ²¹⁰Pb and ¹³⁷Cs dating were performed to reconstruct the MP deposition conditions. The results prove a more frequent accumulation of MP in upper floodplain soils (0–50 cm) deposited by flood dynamics since the 1960s than in subsoils. The first MP detection to a depth of 2 m and below recent (>1960) sediment accumulation indicates in-situ vertical transfer of mobile MP particles through natural processes (e.g., preferential flow, bioturbation). Furthermore, the role of MP as a potential marker of the Anthropocene is assessed. This study advances our understanding of the deposition and relocation of MP at the aquatic-terrestrial interface.
View
... Depending on variables such as the material's buoyancy in relation to seawater, and thus it's levels of exposure to sunlight and warm temperatures, it could either fragment into smaller pieces Lorenzo-Navarro et al., 2018), become captured on seafloors or shorelines (Fulton et al., 2019;Kako et al., 2014), or ultimately collect in one of the world's five major ocean gyres (Eriksen et al., 2014). Estimates of debris pollution sources can be combined with knowledge of these physical processes and calibrated with sea-surface measurements of marine debris to predict the locations and quantities of marine debris in the world's oceans (Eriksen et al., 2014;Jambeck et al., 2015;Law et al., 2014;Lebreton et al., 2019). ...
Automatic detection of Hawai'i's shoreline stranded mega-debris using deep learning-based object detection
Thesis
Full-text available
  • May 2021
Marine debris is a global crisis that negatively affects human health and safety, wildlife, and coastal economies. Geospatial technology has been deployed to map, measure, and model the sources, pathways, and eventual sinks of marine debris, however data are still scarce due to the high cost of conducting fieldwork or manually labeling debris in remote sensing images. Recent advances in artificial intelligence and deep learning have brought rapid automation of tedious tasks to new domains, such as marine debris identification in aerial imagery. This study evaluates deep learning-based object detection to automatically detect, localize, and classify stranded marine debris along 1,200 km of the diverse Hawaiian coastline. Two leading object detection models were tested: Faster RCNN with Inception-ResNet-v2 (FR-IR) and SSD with MobileNetV2 (SS-MN). These two models offer a tradeoff between detection speed and identification precision, with one model performing 5x’s as quickly. The results show that the model with higher computational cost, FR-IR, can detect and classify mega-debris (marine debris larger than 10cm) with a mean average precision (mAP) of 57.4%. The speedier model, SS-MN, detects and classifies mega-debris with an mAP of 36.1%. When viewed in detail the results of our object detection and deep learning classifiers show clear best practices for the advancement of deep learning for real-world management of shoreline stranded marine debris at the large, regional scale.
View
... For example, a study was conducted to accurately identify the types of microplastics isolated from sand or sea water using precision equipment, such as microscopes, electron microscopes, etc. [4]. In addition, a machine learning technique was used to determine the type and number of microplastics isolated from beach sand [5]. In another paper [6], segmentation was performed on images of microplastics separated from sand using U-net [7]. ...
Neural Network Analysis for Microplastic Segmentation
Article
Full-text available
It is necessary to locate microplastic particles mixed with beach sand to be able to separate them. This paper illustrates a kernel weight histogram-based analytical process to determine an appropriate neural network to perform tiny object segmentation on photos of sand with a few microplastic particles. U-net and MultiResUNet are explored as target networks. However, based on our observation of kernel weight histograms, visualized using TensorBoard, the initial encoder stages of U-net and MultiResUNet are useful for capturing small features, whereas the later encoder stages are not useful for capturing small features. Therefore, we derived reduced versions of U-net and MultiResUNet, such as Half U-net, Half MultiResUNet, and Quarter MultiResUNet. From the experiment, we observed that Half MultiResUNet displayed the best average recall-weighted F1 score (40%) and recall-weighted mIoU (26%) and Quarter MultiResUNet the second best average recall-weighted F1 score and recall-weighted mIoU for our microplastic dataset. They also require 1/5 or less floating point operations and 1/50 or a smaller number of parameters over U-net and MultiResUNet.
View
... The use of artificial intelligence in microplastic research is not new; previous studies demonstrated the potential of machine learning in the detection of soil microplastic contamination [59,60], microbeads in wastewater [61] and analysis of spectral images of microplastics [62]. Motivated by the application of ML algorithms on the classification of microplastic in various environmental samples, the possibility of polystyrene particle identification in biological samples was investigated. ...
Label-free identification of microplastics in human cells: dark-field microscopy and deep learning study
Article
  • Oct 2021
The development of an automatic method of identifying microplastic particles within live cells and organisms is crucial for high-throughput analysis of their biodistribution in toxicity studies. State-of-the-art technique in the data analysis tasks is the application of deep learning algorithms. Here, we propose the approach of polystyrene microparticle classification differing only in pigmentation using enhanced dark-field microscopy and a residual neural network (ResNet). The dataset consisting of 11,528 particle images has been collected to train and evaluate the neural network model. Human skin fibroblasts treated with microplastics were used as a model to study the ability of ResNet for classifying particles in a realistic biological experiment. As a result, the accuracy of the obtained classification algorithm achieved up to 93% in cell samples, indicating that the technique proposed will be a potent alternative to time-consuming spectral-based methods in microplastic toxicity research.
View
Explainable Detection of Microplastics Using Transformer Neural Networks
Chapter
  • Dec 2022
Microplastics are environmental contaminants that put marine and aquatic ecosystems at serious risk. Monitoring microplastics is necessary to understand the level of microplastic pollution in our environment. However, the lack of a standard protocol for quantifying and classifying microplastics causes problems in the reliability and comparability of results. Previous literature has employed deep learning models to classify and quantify microplastic polymers with great success, but the ability of these models to classify microplastics from new domains is unanswered. This paper presents an innovative approach to microplastic classification that employs a deep learning approach using a transformer neural network. Our specific contributions are: (1) A novel way to pre-process FTIR spectral data to dramatically increase classification accuracy. (2) Developed a transformer neural network for classifying microplastic polymer FTIR spectra. With the inclusion of a wider range of data, future deep learning approaches will improve the classification and quantification of microplastic polymers, subsequently reducing the costs and labour involved.
View
Automatic quantification and classification of microplastics in scanning electron micrographs via deep learning
Article
Microplastics quantification and classification are demanding jobs to monitor microplastic pollution and evaluate the potential health risks. In this paper, microplastics from daily supplies in diverse chemical compositions and shapes are imaged by scanning electron microscopy. It offers a greater depth and finer details of microplastics at a wider range of magnification than visible light microscopy or a digital camera, and permits further chemical composition analysis. However, it is labour-intensive to manually extract microplastics from micrographs, especially for small particles and thin fibres. A deep learning approach facilitates microplastics quantification and classification with a manually annotated dataset including 237 micrographs of microplastic particles (fragments or beads) in the range of 50 μm–1 mm and fibres with diameters around 10 μm. For microplastics quantification, two deep learning models (U-Net and MultiResUNet) were implemented for semantic segmentation. Both significantly outmatched conventional computer vision techniques and achieved a high average Jaccard index over 0.75. Especially, U-Net was combined with object-aware pixel embedding to perform instance segmentation on densely packed and tangled fibres for further quantification. For shape classification, a fine-tuned VGG16 neural network classifies microplastics based on their shapes with high accuracy of 98.33%. With trained models, it takes only seconds to segment and classify a new micrograph in high accuracy, which is remarkably cheaper and faster than manual labour. The growing datasets may benefit the identification and quantification of microplastics in environmental samples in future work.
View
View
View
Microplastic and tar pollution on three Canary Islands beaches: An annual study
Article
Marine debris accumulation was analyzed from three exposed beaches of the Canary Islands (Lambra, Famara and Las Canteras). Large microplastics (1-5mm), mesoplastics (5-25mm) and tar pollution were assessed twice a month for a year. There was great spatial and temporal variability in the Canary Island coastal pollution. Seasonal patterns differed at each location, marine debris concentration depended mainly of local-scale wind and wave conditions. The most polluted beach was Lambra, a remote beach infrequently visited. The types of debris found were mainly preproduction resin pellets, plastic fragments and tar, evidencing that pollution was not of local origin, but it cames from the open sea. The levels of pollution were similar to those of highly industrialized and contaminated regions. This study corroborates that the Canary Islands are an area of accumulation of microplastics and tar rafted from the North Atlantic Ocean by the southward flowing Canary Current.
View
Imagenet classification with deep convolutional neural networks
Conference Paper
  • Jan 2012
We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 dif- ferent classes. On the test data, we achieved top-1 and top-5 error rates of 37.5% and 17.0% which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implemen- tation of the convolution operation. To reduce overfitting in the fully-connected layers we employed a recently-developed regularization method called dropout that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry
View
A standardized method for sampling and extraction methods for quantifying microplastics in beach sand
Article
  • Sep 2016
Microplastics are ubiquitous in the environment, are frequently ingested by organisms, and may potentially cause harm. A range of studies have found significant levels of microplastics in beach sand. However, there is a considerable amount of methodological variability among these studies. Methodological variation currently limits comparisons as there is no standard procedure for sampling or extraction of microplastics. We identify key sampling and extraction procedures across the literature through a detailed review. We find that sampling depth, sampling location, number of repeat extractions, and settling times are the critical parameters of variation. Next, using a case-study we determine whether and to what extent these differences impact study outcomes. By investigating the common practices identified in the literature with the case-study, we provide a standard operating procedure for sampling and extracting microplastics from beach sand.
View
View
View
Automatic analysis of zooplankton using digitized images: State of the art and perspectives for Latin America
Article
  • Jan 2013
The study of zooplankton communities is very important in the context of the biology, ecology, and conservation of the marine ecosystem. However, research on zooplankton implies a highly costly process in terms of time and effort, requiring trained personnel for identification, counting and measuring of various taxonomic groups, including biomass estimates. At present, there is an increasing interest for implementing new techniques allowing automatic analyses of zooplankton by means of informatics technology and assessment of digitized images. In this work, the state of the art of some methodologies applied to digital images of zooplankton is revised describing the general methods used, such as ZooScan-ZooProcess and ZooImage, and an analysis of their results for identification, a comparison with traditional methods, and recommendations to improve the identification process. Additionally, a brief summary of the experiences with these systems is presented, focused on the analysis of spatial and temporal variability of the zooplankton community by means of its size spectrum and tests for deriving biomass estimates. Finally, a list of capacities and limitations of the methods is discussed using recent literature, as well as the perspectives of development and applications of this new technology can offer to the scientific community of the Latin American region.
View
View
Microplastics in the Ocean
Article
Since their ubiquity in the ocean and marine organisms was first revealed, global concern about microplastics has grown considerably. The North Pacific Ocean and the adjacent marginal seas have high levels of microplastic contamination compared with the global average. This special issue on microplastics was organized by the North Pacific Marine Science Organization to share information on microplastic pollution in the North Pacific region. The special issue highlights high levels of contamination in the North Pacific both on shorelines and at the sea surface. Particularly high levels of contamination were reported on the western and southern coasts of Korea. Sources, including sewage discharge, aquaculture, and shipyards, were implicated. With the direction and energy of surface winds and currents have an important influence on shoreline patterns of distribution. The special issue also demonstrates potential for ingestion of microplastic by small planktonic organisms at the base of the food chain. A wide range of chemicals are associated with plastic debris and concerns are expressed about the potential for these chemicals to transfer to biota upon ingestion. As an introduction to the topic, this paper provides a brief background on microplastic contamination, highlights some key research gaps, and summarizes findings from the articles published in this issue.
View
Microplastics in sediments: A review of techniques, occurrence and effects
Article
  • Jun 2015
Microplastics are omnipresent in the marine environment and sediments are hypothesized to be major sinks of these plastics. Here, over 100 articles spanning the last 50 year are reviewed with following objectives: (i) to evaluate current microplastic extraction techniques, (ii) to discuss the occurrence and worldwide distribution of microplastics in sediments, and (iii) to make a comprehensive assessment of the possible adverse effects of this type of pollution to marine organisms. Based on this review we propose future research needs and conclude that there is a clear need for a standardized techniques, unified reporting units and more realistic effect assessments. Copyright © 2015 Elsevier Ltd. All rights reserved.
View